國立虎尾科技大學 |

Dimensionality Reduction of Embedded Boundary Models for Nonlinear Fluid-Structure Interaction.

Record Type:	Language materials, manuscript : Monograph/item
Title/Author:	Dimensionality Reduction of Embedded Boundary Models for Nonlinear Fluid-Structure Interaction./
Author:	Youkilis, Noah Ben.
Description:	1 online resource (172 pages)
Notes:	Source: Dissertations Abstracts International, Volume: 85-06, Section: B.
Contained By:	Dissertations Abstracts International85-06B.
Subject:	Aircraft. -
Online resource:	click for full text (PQDT)
ISBN:	9798381026627

Dimensionality Reduction of Embedded Boundary Models for Nonlinear Fluid-Structure Interaction.
Youkilis, Noah Ben.

Dimensionality Reduction of Embedded Boundary Models for Nonlinear Fluid-Structure Interaction. - 1 online resource (172 pages)

Source: Dissertations Abstracts International, Volume: 85-06, Section: B.

Thesis (Ph.D.)--Stanford University, 2023.

Includes bibliographical references

Embedded (sometimes immersed) boundary methods (EBMs) for CFD and fluidstructure interaction (FSI) - also known as immersed boundary methods, Cartesian methods, or fictitious domain methods - are the most robust solution methods for flow problems past obstacles that undergo large motions, deformations, shape changes, and/or surface topology changes. They can also introduce a high degree of automation in the task of mesh generation and significant flexibility in the meshing of complex geometries. However, despite massive advancements in compute power, the application of CFD- or FSI-based EBMs to the simulation of many realistic engineering systems can still take days or weeks on many cores of a supercomputer, rendering the use of such models in many-query applications challenging, impractical or simply infeasible.The application of projection-based model order reduction (PMOR) to nonlinear CFD and FSI models has been shown in recent years to produce low-dimensional projection-based reduced order models (PROMs) which can be queried in real-time, without sacrificing the accuracy of the underlying high-fidelity, high-dimensional computational model. However, the application of PMOR to EBMs for CFD and FSI remains a challenge, primarily because EBMs dynamically partition the computational fluid domain into real and fictitious subdomains. Consequently, despite their myriad benefits, EBMs hamper the acceleration of the solution of computationally intensive problems by these types of physics-based surrogate models.This dissertation addresses several issues stemming from the real/fictitious partitioning of the computational fluid domain to pave the way for the PMOR of EBMs.Specifically, it motivates, presents, and illustrates a computational framework for constructing EBM-PROMs that incorporates state-of-the-art methodologies for mitigating computational bottlenecks associated with strong nonlinearities in the underlying high-dimensional model. This framework includes a weighted low-rank matrix factorization approach for computing suitable reduced-order bases; and adaptations of the energy-conserving sampling and weighting hyperreduction procedure as well as the piecewise-affine approximation method that account for the real/fictitious dynamic partitioning of the nodes of the embedding CFD mesh. It also employs an adaptive sampling approach for choosing an optimal set of parameter points on which to train a parametric EBM-PROM, based on a novel error indicator. Finally, the performance of the proposed computational framework for the PMOR of embedded boundary models is investigated for two representative, unsteady, nonlinear, turbulent flow problems that necessitate the use of an EBM; as well as two prototypical steady-state, shape-parametric studies of relevance to MDAO.

Electronic reproduction.
Ann Arbor, Mich. :
ProQuest,
2024

Mode of access: World Wide Web

ISBN: 9798381026627Subjects--Topical Terms:

580976
Aircraft.
Index Terms--Genre/Form:

554714
Electronic books.

Dimensionality Reduction of Embedded Boundary Models for Nonlinear Fluid-Structure Interaction.
LDR:04157ntm a22003737 4500 001 1141805
005 20240414211523.5
006 m o d
007 cr mn ---uuuuu
008 250605s2023 xx obm 000 0 eng d
020 $a 9798381026627
035 $a (MiAaPQ)AAI30742192
035 $a (MiAaPQ)STANFORDzm574wr0384
035 $a AAI30742192
040 $a MiAaPQ $b eng $c MiAaPQ $d NTU
100 1 $a Youkilis, Noah Ben. $3 1465798
245 1 0 $a Dimensionality Reduction of Embedded Boundary Models for Nonlinear Fluid-Structure Interaction.
264 0 $c 2023
300 $a 1 online resource (172 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Dissertations Abstracts International, Volume: 85-06, Section: B.
500 $a Advisor: Farhat, Charbel;Alonso, Juan;Kochenderfer, Mykel.
502 $a Thesis (Ph.D.)--Stanford University, 2023.
504 $a Includes bibliographical references
520 $a Embedded (sometimes immersed) boundary methods (EBMs) for CFD and fluidstructure interaction (FSI) - also known as immersed boundary methods, Cartesian methods, or fictitious domain methods - are the most robust solution methods for flow problems past obstacles that undergo large motions, deformations, shape changes, and/or surface topology changes. They can also introduce a high degree of automation in the task of mesh generation and significant flexibility in the meshing of complex geometries. However, despite massive advancements in compute power, the application of CFD- or FSI-based EBMs to the simulation of many realistic engineering systems can still take days or weeks on many cores of a supercomputer, rendering the use of such models in many-query applications challenging, impractical or simply infeasible.The application of projection-based model order reduction (PMOR) to nonlinear CFD and FSI models has been shown in recent years to produce low-dimensional projection-based reduced order models (PROMs) which can be queried in real-time, without sacrificing the accuracy of the underlying high-fidelity, high-dimensional computational model. However, the application of PMOR to EBMs for CFD and FSI remains a challenge, primarily because EBMs dynamically partition the computational fluid domain into real and fictitious subdomains. Consequently, despite their myriad benefits, EBMs hamper the acceleration of the solution of computationally intensive problems by these types of physics-based surrogate models.This dissertation addresses several issues stemming from the real/fictitious partitioning of the computational fluid domain to pave the way for the PMOR of EBMs.Specifically, it motivates, presents, and illustrates a computational framework for constructing EBM-PROMs that incorporates state-of-the-art methodologies for mitigating computational bottlenecks associated with strong nonlinearities in the underlying high-dimensional model. This framework includes a weighted low-rank matrix factorization approach for computing suitable reduced-order bases; and adaptations of the energy-conserving sampling and weighting hyperreduction procedure as well as the piecewise-affine approximation method that account for the real/fictitious dynamic partitioning of the nodes of the embedding CFD mesh. It also employs an adaptive sampling approach for choosing an optimal set of parameter points on which to train a parametric EBM-PROM, based on a novel error indicator. Finally, the performance of the proposed computational framework for the PMOR of embedded boundary models is investigated for two representative, unsteady, nonlinear, turbulent flow problems that necessitate the use of an EBM; as well as two prototypical steady-state, shape-parametric studies of relevance to MDAO.
533 $a Electronic reproduction. $b Ann Arbor, Mich. : $c ProQuest, $d 2024
538 $a Mode of access: World Wide Web
650 4 $a Aircraft. $3 580976
650 4 $a Construction. $3 1254179
650 4 $a Partial differential equations. $3 1102982
650 4 $a Vortices. $3 1413589
650 4 $a Fluid-structure interaction. $3 681130
650 4 $a Neural networks. $3 1011215
650 4 $a Design. $3 595500
650 4 $a Fluid mechanics. $3 555551
650 4 $a Mechanics. $3 527684
655 7 $a Electronic books. $2 local $3 554714
690 $a 0389
690 $a 0800
690 $a 0543
690 $a 0204
690 $a 0346
710 2 $a ProQuest Information and Learning Co. $3 1178819
710 2 $a Stanford University. $3 1184533
773 0 $t Dissertations Abstracts International $g 85-06B.
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=30742192 $z click for full text (PQDT)